Friday, November 03, 2006

The art of coding HDTV

I have heard quite a few figures cited for the efficiency of the H.264/MPEG-4 advanced video coding (AVC) standard in encoding high definition TV. What I didn't understand were the issues dictating the progress being made.

Having spoken to video coding specialists Harmonic and Tandberg Television I now have a better idea (see table below for actual coding results). Not surprisingly, the main issues are hardware and the vendors' algorithmic expertise.

With regard algorithms, the AVC standard has many coding techniques but it doesn't say which should be used when. That expertise is built up over time and captured by the vendors in software with each new platform generation.

As for hardware, Harmonic's first generation HD encoder requires that a video frame be split into slices, each processed alone because of bottlenecks in moving the data between processors. Now, with its second generation design, a whole video frame can be processed improving the effectiveness of the compression techniques.

Both Harmonic and Tandberg claim over a 30 percent coding improvement using their latest designs.

Why is this important? Well, a lower bit rate means more channels per broadband link or a greater percentage of households that can be reached with HD services using existing DSL infrastructure. Harmonic also believes there is a further 25 percent improvement to be gained in the coming few years.

For background material on video coding, click here.
For an article on coding, click here


MPEG-2

AVC

Standard definition TV

2 to 2.5Mbit/s (cable)

4 to 4.5 Mbit/s (Europe)

1.5Mbit/s to 2Mbit/s

High Definition TV

9.5Mbit/s (cable)

13 to 14Mbit/s (Constant bit rate over DSL. Satellite and cable use variable bit rate schemes that allows statistical multiplexing of the channels)

1st generation encoders: 9.1Mbit/s

2nd generation: 6.2Mbit/s (cable)

7.5 to 8Mbit/s (constant bit rate over DSL)

No comments: